Compared to current practice with whole-body-gradient coils, an advanced head gradient coil promises over twice the gradient strength (for higher resolution) and at least three times higher slew rates (for faster imaging sequences), with an order of magnitude reduction in nerve stimulation and greatly reduced acoustic noise. The potential usefulness of a high-performance head gradient coil has been widely recognized for more than a decade, and many groups around the world have each put millions of dollars into attempting to developing a satisfactory solution, but to limited avail. The enormous increases in S/N in head MR RF coils reported over the past year serve to further strengthen the justification for a high-performance dedicated head gradient coil. Preliminary simulations indicate that an optimal combination of a number of improvements in our pre- vious crescent-coil approach will permit the realization of a practical, robust, head gradient coil with dramatically reduced acoustic noise, higher gradient strength, and substantially reduced manufacturing cost. The higher gradient performance is especially needed at higher fields for improved brain functional imaging (fMRI). The primary justification in clinical applications may be the improved S/N that will be obtained in Diffusion Weighted MRI techniques. Next-generation ultra-fast hyperpolarized 13C techniques will also benefit. Moreover, a dedicated head MR scanner would be less expensive than a whole body scanner while providing greatly improved performance for head and extremity applications in patients of all ages, and for whole-body imaging of infants. Such a system could supplant x-ray CT in more cases. Based on preliminary simulations, the following performance is expected: 200 mT/m gradient at 900 A; slew rate of 520 T/m/s at 1400 V; 25 cm Region of Uniformity imaging diameter with 10% rms non-uniformity; residual eddy current effects under 0.4%; acoustic noise below 100 dB for a typical EPI brain sequence in a 3 T field; and mechanical robustness suitable for operation with no current de-rating at fields at least up to 8 T. The RT shims will provide major corrections through Z4, including dynamic Z0 and Z2. The Phase I will complete detailed simulations and balanced optimization of all parameters. A prototype coil will be constructed, and preliminary tests will be completed. More extensive testing will be carried out during Phase II at another institution, followed by revisions based on lessons learned and construction and testing of the pre-production version. [unreadable] [unreadable] [unreadable]